Vibration diagnosis of a gearbox by wavelet bicoherence technology

Abstract

Gearboxes are critical elements of mechanical systems that are widely used in aerospace, energy generation, land and naval applications. The early detection of changes in the technical condition of this equipment is of great importance for the optimisation of maintenance costs. Vibration signal components resulting from the presence of the developing faults of meshing gears contain the information that, once extracted from the signal, may allow for a reliable estimation of the technical condition of the meshing gears. Wavelet bicoherence (WB)-based technology has been used to obtain the signal feature characterising the phase relationship between the signal components generated by gear faults in the selected frequency bandwidths. In previous research, WB has been successfully applied to the detection of artificiallycreated gearbox faults. This paper will present the application of WB in the detection of naturally-developing gear faults.

Novel Higher-Order Spectral Cross-Correlation Technologies for Vibration Sensor-Based Diagnosis of Gearboxes

“…The plot below ( Figure 5 ) shows the estimation of the micro-pitting progress on the pinion teeth, which showed damage [ 27 ].…”

“…Taking into account the above two considerations, the combination of the HOS with the wavelet transform, i.e., the wavelet HOS [ 22 , 23 , 24 , 25 , 26 , 27 ], is superior over the classical HOS for early local fault diagnosis in gearboxes and bearings. The higher effectiveness of the wavelet HOS over the classical HOS is shown in the pioneering research work [ 22 ].…”

“…The locally averaged and instantaneous WB-based technologies are applied to rolling element bearings fault detection and localisation [ 26 , 34 , 48 ]. The WB is applied for detection of pitting on a gearbox in [ 27 ]. For the proposed approach, the selection of frequency bands is based on evaluation of the averaged WB modulus maps.…”

Novel Instantaneous Wavelet Bicoherence for Vibration Fault Detection in Gear Systems

“…Because of non-linearity, which characterises a variety of gear systems, most vibration transients, which are related to gear faults, create statistically dependent characteristic spectral components. Therefore, higher order spectra (HOS), capable of detecting their presence in vibration signals, have been broadly used in the field of machinery fault detection, e.g., [33][34][35][36], and in various domains of science, e.g., [37,38]. In reference [33], the wavelet HOS and HOS integrated features are used for early-stage gear fault detection.…”

“…The research presented in reference [34] uses the classical Fourier HOS for the detection of distortions of the intake airflows of an aircraft engine, known as an intake separation. Reference [35] presents, for the first time, the application of wavelet bicoherence to the detection of naturally developed early-stage gearbox pinion faults. In reference [36], the bispectrum and the trispectrum are used for the detection of relations between multiple spectral components, which appear in analysed signals as a result of fault-related nonlinear behaviour.…”